Citation

Breault KD. Am. J. Sociol. 1986; 92(3): 628-656.

Copyright

(Copyright © 1986, University of Chicago Press)

DOI

10.1086/228544

PMID

unavailable

Abstract

With a set of relevant controls, including population change, income, urbanity, unemployment, and female labor-force participaion, Durkheim's religious and family integration hypotheses are tested with new and highly reliable church membership rates and similar divorce rates. Eight data sets are analyzed at two levels of analysis, state and county: 50 states at six different times betwen 1933 and 1980 and 216 counties in 1970 and 1980. The religous and family integration variables are generally supported at both levels of analysis and across the 47 years of the study. Contrary to previous studies, support is found for Durkheim's thesis that Catholics have lower suicide rates than non-Catholics. Strong Catholic/non-Catholic suicide differences are found even when the analysis is extended to 414 counties in 1970 and 1980.